DE10113880B4 - Method for compressing and decompressing video data - Google Patents

Abstract

A method for compressing and decompressing video data consisting of an array of individual image points (pixels). Each pixel has a pixel value that changes with time and that describes the colour or luminosity information of the pixel. A priority is allocated to each pixel and the pixels are then stored in a priority array according to this priority allocation. The array contains at each moment in time the pixel values that have been classified according to the priority allocation. The pixels and the pixel values that have been used to calculate the priority allocation are transmitted or saved according to the priority allocation. A pixel receives a high priority if the differences in relation to its neighbouring pixel are great. For the reconstruction process, the current pixel values in each case are reproduced on the display. The pixels that have not yet been transmitted are calculated from the pixels that have already been transmitted

Description

The invention describes a method for compressing and decompressing video data

Videos generate large amounts of data. In order to effectively transfer and store this amount of data, it makes sense to compress them.

According to the current state of the art become video signals in a rapid succession of frames recorded and reproduced. For television (PAL standard), these are 25 frames per second or 50 fields. For digital recordings that's up to about 30 frames per second. Each picture is in lines disassembled, and initially transmitted sequentially.

Previous compression methods are essentially based on reducing the resolution of the Color depth and reducing the number of frames per second. In digital compression, e.g. the MPEG process, are instead of complete images essentially the difference images, i.e. the differences of the individual pixels (pixels) compared to the previous image, instead of the full images transferred. The latest standard for video coding is MPEG4.

MPEG is the abbreviation for "Motion Pictures Expert Group". From this group were and are file formats and procedures for space-saving Compress and save video or multimedia data (video, Image and sound data) in high quality sets. The MPEG standard is now divided into MPEG-1, MPEG-2, MPEG-3 and MPEG-4, whereby the MPEG-3 standard has now been integrated into MPEG-2.

To the huge amount of data from films process with "normal" computers and to be able to transport only the changes to the previous picture stored. The MPEG format periodically saves from typically twelve Images so-called intra-frames; these are JPEG compressed single images. The images between these I-frames are, if possible not completely filed. Rather, MPEG stores how to go through them Moving parts from previous or following pictures win back can. Predictive frames and B-frames (bi-directional frames) are also used for this purpose. But since this never works perfectly, the remaining ones per image are added Deviation still saved JPEG-coded. With this method you can the data overhead for reduce a video movie by about 99%. The possible compression goes up 200: 1st

MPEG-1 is designed for smooth video playback. The MPEG-1 compression or decompression was original a hardware dependent Method. However, thanks to the fast processors, software decompression is also possible.

The main difference between MPEG-1 and MPEG-2 is that MPEG-2 is better with that at Bypass interlaced television can. The secret of MPEG-2 is compression at the highest quality level, so footage almost 1 to 1 in studio quality can be edited and edited. Consistently established MPEG-2 became a standard. With pure I-frame coding let yourself Use MPEG-2 even in editing mode.

The part of the MPEG-3 standard that for high Definition of TV quality (HDTV) was provided, has meanwhile become in the MPEG-2 standard implemented. MPEG-4 is a further development of the MPEG-2 format and has been in development since 1996. Although MPEG-4 was originally called a coding standard for audiovisual Data with very low bitrate was intended for development far more than just streaming linear media data in internet and wireless applications. MPEG-4 e.g. efficient Mechanisms for compressing and distributing interactive media content ready. In addition, MPEG-4 has 3D potential for artificial Visualize intelligences or display avantars, e.g. in the context of video conferences.

The compression rate is at MPEG-4 higher than with MPEG-2, whereby "Sprites" compresses better can be because the coding mechanism for it much more time available stands. It may even be possible to switch to wavelets. The scripting language enables it is possible to perform operations such as "move" much faster in a few bytes than allow the digitized compressed form of the same operation would. With the help of these "sprites" can be any contoured still images over moving images are pushed.

From the DE 692 26 825 T2 ( EP 0 521 662 B1 ) a method for processing image data is known which consists of an array of individual pixels. Each pixel of a video image is compared to a preselected group of neighboring pixels, and priority values are determined in such a way that local area maxima and local area minima are displayed. The method is based on the use of autocorrelation.

The DE 696 10 987 T2 ( EP 0 845 191 B1 ) relates to a method for compressing image data, which is essentially based on a prediction of image data from already scanned image data.

The DE 695 19 462 T2 ( EP 0 705 027 B1 ) discloses a method for data reduction of color images by means of color tables and summary of image areas.

The DE 694 25 047 T2 ( EP 0 685 137 B1 ) discloses an adaptive variable length encoding and decoding method for image data using a plurality of variable length encoding tables.

The DE 693 24 538 T2 ( EP 0 588 476 B1 ) relates to an image data compression method which is based on a splitting of the image frequencies with subsequent quantization of the frequency components.

The DE 195 41 457 C1 teaches a method for encoding a video data stream of a video sequence consisting of picture blocks, which is based on a performed motion estimation for each picture block.

The object of the invention is in the creation of a method for compressing video data, which a simple and flexible adaptation to different transmission rates or bandwidth, image resolutions and display sizes allowed.

This task is due to the characteristics of claim 1 solved.

Advantageous configurations and Developments of the invention are specified in the subclaims.

A parallel is preferably carried out Processing the video information in the video recording chip. The parallel Processing serves the purpose of first the most important pixels to determine and this according to a prioritization in a priority array store. This array contains at any time, the pixel values sorted according to the prioritization. According to the prioritization, these pixels and those for the calculation the pixel values used in the prioritization, transmitted or stored. A pixel gets high priority if the differences are too its neighboring pixels are very large.

They are used for reconstruction current pixel values shown on the display. The not yet transferring Pixels become those that are already transmitted Pixels calculated.

According to the computing power, the transmission bandwidth and the size of the display can be different Methods for calculating the pixels not yet transferred are used become. Stands a very big one Bandwidth available a simple linear interpolation can be performed. If only a very small bandwidth is available, this can happen during transmission of the prioritized pixels are taken into account.

Through the history of the transferring Pixels can Identifies objects, and thus a motion estimate of them Objects carried out become.

The basic idea of the process is based on the prioritized pixel storage or transmission. When saving or the video transmission have to furthermore the temporal and positional (within the image array) dependencies of the individual pixels or pixels grouped in pixel groups considered become.

An extremely high data compression to achieve the pixel groups that have the highest priority are transmitted, and have not yet been broadcast have been. The areas i.e. the pixel values between the already transmitted pixel groups from those already transmitting Pixel groups calculated, e.g. through interpolation. At higher resolution (larger image array) the achievable compression factor increases, because with natural Recordings of larger areas mostly have a well predictable (even) color gradient, e.g. blue sky.

It should also be noted that transfer the exact pixel values become. If necessary, this procedure allows lossless transmission the video information.

The restoration or reconstruction the video data is based on estimates similar to human Viewing patterns. Man perceives stimuli, but interpretation what he recognizes in this picture only takes place in his brain. The stimuli correspond to those transferred Pixel groups, the interpretation corresponds to filling in the surfaces between those not yet transferred Pixel groups.

To achieve this, additional Arrays can be created. Among other things, these are an array in which the information is available from which pixel groups the pixel value was determined at the current position. Further information can when these values were calculated, from which pixel groups this is calculated or transferred have been. Also an assessment of the Accuracy of the values (e.g. calculation from directly neighboring pixels, small variance of the pixels on which the calculation is based) as additional Information can be evaluated.

The method described allows a much easier adaptation of the video data stream to different Display sizes and Resolutions.

Another advantage is that by the described way of encoding the video is not automatically determined with which algorithms the video has to be decoded. This is through transmission of the priority values reached, which contrary to others Procedure did not experience averaging. Manufacturers have with it the possibility Low to high cost devices too develop and different algorithms from the competition emphasized.

The targeted massively parallel Processing of the video data in a specially developed chip allows to use extremely low clock speeds, which is cheap on affects electricity consumption.

By prioritizing certain Areas of the video (e.g. lips for news speakers) with a higher priority and consequently transferred to a better resolution become.

The procedure allows from the Stream of video for different end devices the optimal partial data flows filter out without taking this into account when recording video must become.

An embodiment of the invention is explained in more detail using the drawing figures. From the drawings and their description, further features, advantages and applications the invention. It shows:

1 : Representation of an image array of 20 × 21 pixels;

2 : Representation of different forms of pixel groups;

3 : Image array with moving object at a time t1;

4 : Image array with moving object at a time t2;

5 : Image array with moving object at a time t3;

6 : Newly generated image array with inserted pixel groups in the image corner;

7 : Filling in the areas between the already inserted pixel groups;

8th : Insert further pixel groups and fill in the areas in between.

Below is an example For example, compression and decompression of a video signal described.

The following assumptions are made met:

A currently common video signal is available as a video source (e.g. PAL or NTSC). The video information can be read out with a commercially available electronic device (eg frame grabber card). In order to illustrate the method, a minimized image array with a width of 20 pixels and a height of 21 pixels is used in the following ( 1 ). Each pixel of the array is represented by a 32 bit value (pixel value). The 32 bits are divided into 4 values (transparent, red, green, blue), each with 8 bits. The position of the pixels is determined by an integer. The image array is in the in 1 counted from 0 to 419 as shown. The number within each box corresponds to the position of the associated pixel. There is a UDP (User Datagram Protocol) connection between the source and the sink. The compressed video data is then sent via these.

The compression of the Video signal is as follows:

The process is based on the fact that constantly the individual pixels of the video signal are prioritized, whereby the pixels are stored in an array according to their prioritization become. This array contains at any time, the current pixel values sorted by priority. On Pixel gets high priority when the differences to its neighboring pixels are very large. The pixel is used along with its for the calculation Neighboring pixels combined into a pixel group. Corresponding These pixel groups are transmitted or stored in the prioritization.

Read in image array

The frame grabber always has the current image in its image array as in 1 shown, for example, can be a 20 × 21 pixel image array. Each pixel is defined by its position (0 to 419) and its pixel value (color or brightness value).

Set pixel groups

It was previously determined which neighboring pixels form a pixel group. P0 denotes the pixel that indicates the position of the pixel group and for which the priority is calculated. The relative position of the remaining pixels, for example p1-p4, of a pixel group to the reference pixel p0 results from the type (shape) of the pixel group used. In 2 some possible forms of pixel groups are shown as examples. Both symmetrical and asymmetrical pixel groups can be formed with respect to the reference pixel p0. Which type of pixel group is used depends, among other things, on the type of image material and the desired compression rate. As a rule, the compression factor to be achieved is greater, the more pixels a pixel group comprises. For coding and decoding, ie compressing and decompressing the Vi The same form of pixel groups must be used.

priority values determine

For each pixel p0 of a pixel group now becomes the priority in relation calculated on its pixel group. Each pixel is 0 - 419 of the Once to the reference pixel p0. It is provided according to the invention that the calculation of the priority values the pixel as much as possible is carried out as part of parallel data processing. It is optimal if the calculation of the priorities of all pixels of the image done simultaneously. That it is a video signal the priority values all pixels constantly recalculated because the image content is constantly changing. The multitude of pixel groups, especially those with lower priorities, but will with a high probability not to change.

It can be used to calculate the priority give different calculation methods. A linear example is used here as an example Method used.

To do this, the individual pixel values P0, P1, P2, P3 and P4 of a pixel group in their color components red, Green and Disassembled blue. Each of these color values is represented by 8 bits. For every Color of each pixel P1-P4 a color difference value is now determined in relation to P0, e.g. P0_rot-P1_rot, P0_rot-P2_rot, ..., P0_blau – P4_blau. The absolute color difference values are added and by the number the colors and number of pixels viewed. The result is a priority value for the considered pixel group. This priority value is the higher, ever the color values of the individual pixels of the pixel group are more different are.

Other methods of determining the priority value are the use of gray values or the maximum value of a color difference one color. Because the priority value later not broadcast yourself or stored, the method for determining the priority value has none direct influence on the decoding.

This prioritization ensures that image areas that have a large color or contrast change have, such as Edges, given high priority, and relatively consistent Image content, e.g. blue sky, a low one.

priority values sort by

In this step the priority values in size sorted in descending order. The sorting is done after determining each new priority value For each At this point in time, one has a list of pixel groups arranged according to priorities which is sorted in descending order. It is desirable that the corresponding Image recorders (CCD chips) are developed, which immediately one according to priorities Provide an ordered list.

The image to be compressed becomes direct in principle, recorded with a CCD camera or a scanner the possibility, from the existing image processing in the camera / scanner Microchip to directly get an array sorted by priority. Thus, a significant part of the computational effort when compressing saved.

Update the priority values

In contrast to still images (e.g. photographs), there is a constantly changing change in priority of the pixel groups with video information, e.g. with camera panning or moving objects. To clarify this is in the 3 to 5 a video image array is shown at different times t1 to t3, an object shifting from right to left.

According to 2 the image at time t1 contains an object representing the pixels 156 . 157 . 176 . 177 . 191 - 197 . 211 - 217 . 231 - 237 . 256 . 257 . 276 . 277 fills.

To calculate the priorities of the pixels ( 0 - 419 ) of the picture becomes the in 2 thick bordered shape of the pixel group (bottom left) is used.

Tabelle 1 This results in a priority distribution of the pixels, as is shown in the further description of the method as an example in Table 1 at time t1. The table contains only the number of the reference pixel (p0) of a pixel group. Those pixel groups that are located in the edge area of the object and for which the respective reference pixel (p0) has the greatest difference to the other pixels of the pixel group are given the highest priority A. The pixel groups whose reference pixels have a smaller difference to the other pixels of the pixel group have a medium priority B and those pixel groups whose reference pixels have no difference to the other pixels of the pixel group have the lowest priority C.

Table 1

In the course of the compression, the pixel groups with priority A are now first transmitted or stored, then the pixel groups with priority B and finally the pixel groups with priority C: since the object is now moving and in relation to 3 in the 4 and 5 takes a different position, the priorities of the individual pixel groups change. The priority list is updated and the transfer of the current pixel groups with the highest priority is continued immediately.

The newly calculated priorities of the pixel groups for the times t2 ( 4 ) and t3 ( 5 ) are shown in Table 1.

A possible compressed transmission of a video signal according to Table 1 could thus look as follows:
Time t1: Pixel groups with the highest priority A are transmitted: 175 . 255 . 231 . 191 . 156 . 157 . 277 . 276 . 177 . 197 . 217
New priorities are recognized at time t2. Other pixel groups receive highest priority A. The transfer of the new priority A is continued:
189 . 173 . 154 . 155 . 274 . 275 . 253 . 229
This is followed by the transmission of the pixel groups with priority B:
175 . 195 . 215 . 235 . 255 . 190 . 191 . 192 . 193 ...
New priorities are recognized again at time t3. Other pixel groups receive highest priority A. The transfer of the new priority A is continued:
187 . 171 . 227 . 251 . 152 . 153 . 272 . 273

An additional array is used to check which Pixel groups already transferred have been. If a group of pixels has already been transferred, you do not need it broadcast a 2nd time provided their priority not changed in the meantime Has. Certain image regions, e.g. Faces, can be recognized and preferably transmitted become. additionally can from the recipient certain groups of pixels are also requested (e.g. upon detection of transmission errors due to incorrect CRC check). Pixel groups requested in this way can then a high priority received so that you can transfer immediately becomes.

Save / transfer pixel groups

The terms "save" and "transfer" are as follows used synonymously. First some parameters of the Saved or transmitted video image. Listed as an example for example:

• - image width (in pixels)
• - image height (in pixels)
• - Used Shape of the pixel group (not necessary if only one shape is standardized is)

Then the individual pixel groups according to their priority saved or transferred, i.e. High priority pixel groups are saved first (and later also read out first).

The position value of the Reference pixel p0 of the pixel group is stored. Then be the pixel values P0, P1, P2, P3, P4 are stored.

Example:

Position value P0, pixel values P0, P1, P2, P3, P4; Next Position value P0 (with the same or lower priority), pixel values P0, P1, P2, P3, P4, ..., next Position value P0 (with lowest priority), pixel values P0, P1, P2, P3, P4.

The saving can be done by different Methods that are only mentioned here as examples are optimized become.

There can be a run length coding of the pixel groups. For example, if in an image area no red components occur, this can only be done instead of 8 bits (red) with e.g. 2 bits transmitted or the number of leading zeros can be used.

Common compression methods, e.g. Zip format, can be used.

By setting a limit for the Prioritization can ensure a certain quality. For example, can a limit for the pixel difference values are defined below which the assigned one Pixel group always gets the lowest priority value.

First the 4 pixel groups of the Transfer key points, you get the largest possible area calculated with a few pixel groups.

Reconstruction (decompression) the video data

Generate new image array

When reconstructing the compressed video data, an image array is initially comparable to the representation in 1 generated. For this purpose, the characteristics of the image are read in and evaluated. Examples are the image width, image height and shape of the pixel group used for compression. If the image height and image width do not match between the original image and the desired display (e.g. limited PDA display or high-resolution screen), it must be scaled accordingly. For this purpose, conversion factors are first determined (image width_original / image width_display and image height_original / image height_display). These factors can be used to convert the position value from the original image to the position value of the new display.

Insert pixel groups

As in 6 is shown, these are now read in according to the order of the prioritized pixel groups. For example, the first four pixel groups with the highest priority are entered in the image array. In 6 these are the pixel groups at the corners of the image. The position of the reference pixel p0 of the respective pixel group is indicated by the fields highlighted in black 21 . 38 . 381 respectively. 398 certainly. This position value (p0) is available as an integer value in the saved file. The dark gray-gray pixel values (p1 – p4) belonging to the respective pixel group can then be entered in the new image array. The intermediate, light gray marked pixel values can then be calculated from the dark gray and black marked fields. For the calculation, the known pixel values are first broken down into their components red, green and blue.

The mean value of each color is then calculated, e.g. pixels ( 22 ) = (Pixel ( 2 ) + Pixels ( 21 ) + Pixels ( 42 )) / 3).

Fill in areas

Now the existing pixel groups are connected with each other by lines. This is in 7 shown. The result is triangles, the corners of which are defined by the corresponding pixel groups. This should be exemplified by the line between pixel position 2 and pixel position 17. The color gradient of the line is calculated based on the color values of pixels 2 and 17. First the number of pixels between these two positions is determined, in example 14. Then the color difference is determined for each color (red, green, blue), eg color value at position 2 = 2 ; Color value at position 17 = 30 results in color difference of 28 ). An increase in color value per pixel - from the pixel 2 to the pixel 17 - is then calculated from the color difference / number (in the example 28/14 = 2 ).

The remaining areas are drawn filled with horizontal lines, e.g. from position 63 to position 74, from position 82 to position 93, etc. Again, a preliminary one Gradient between the dots calculated as given above,

How 8th shows, every further added pixel group results in further triangles which can be filled in accordingly. After first using the 4 corner points ( 21 . 38 . 398 . 381 ) has been filled in, the resolution can now be refined with each additional pixel group. The add towards the pixel group 87 leads to 4 triangles with the reference points ( 21 . 38 . 87 ), ( 21 . 87 . 381 ), ( 381 . 87 . 398 ), ( 398 . 78 . 38 ). Is now within such a triangle, e.g. 87 . 381 . 398 , another group of pixels ( 247 ) inserted 3 new triangles ( 247 . 381 . 398 ), ( 247 . 87 . 381 ) and ( 247 . 87 . 398 ). Each new pixel group creates 3 new triangles that can be filled in. The more pixel groups are inserted, ie the more triangles are formed, the closer the calculated color gradient comes to the actual color gradient of the image. Since only new triangles are created from now on, optimized methods can be used for the calculations. Furthermore, each can 3 newly created triangles are calculated in parallel to increase the processing speed.

Additional possibility of parallelization arises when new pixel groups in different regions of the picture added become.

With the process steps described above it was assumed that the image content did not change in the meantime changed Has. change the picture content, then the priorities of the individual pixel groups redistributed and the current pixel groups with the highest priority transmitted. It changes just the order of the just transferred and in the picture pasted Pixel groups. On the principle of reconstruction of the Image changes nothing, however.

To the temporal changes However, additional arrays can take the image content into account (with the size of the image array) be generated. these can Information contains about the

• - Time, i.e. when was a pixel value last calculated or transferred
• - Basis of calculation. Which transferred Pixels were used to calculate the pixel value
• - probability / accuracy. A pixel value was transferred or calculated; if it was calculated how large is the variance of the pixel groups from which the new value was calculated?
• - deviation of already calculated pixel values with then transferring pixel values

Image regions can then be made from these sizes determine in which often Pixel group changes occurrence. Adjacent pixel groups, or entire areas are usually similar changes, e.g.

Changes in brightness, changes in color subject. By evaluating these changes in all Determine rule objects and their dynamic behavior, e.g. object that moves in the video. Uniform changes across the board Image array can for example to point out a camera pan. Will this information e.g. with the help of learnable evaluated neural networks, estimates can be made very easily via the Pixel values from not yet transferring Make pixel groups. If these estimates are correct, pixel groups can be identified, the particular influence on changes own objects. These groups of pixels will come from the source again asked, it is possible To precisely determine object movements with only a few pixel groups and predict. In practice it means that although only a low one Bandwidth available stands, low delay times occur, which are significantly lower than with frame-based methods. The evaluation of the additional in the receiver generated arrays also allow good object recognition.

Depending on the available resources can in addition to pure prioritization by the color values of neighboring ones Pixels also dependencies the position of the prioritized pixel groups can be used.

An application case should clarify this. If you look at a horizon at sea, it looks like one horizontal line. It is expected that the priority values each group of pixels along this horizon are approximately the same. In in this case have the most widely spaced points the greatest meaningfulness of the horizon line. By transmission the outermost left and extreme right pixel groups of the horizon it is already possible to use this to reconstruct again.

Another way of prioritizing lies in the higher rating certain areas of the image. Such an image area can be used for Example faces. Although faces on vacation videos sometimes make up only a small percentage of the entire image, they are usually the focus when looking at them. Such a human Visual behavior can be improved by prioritizing the pixel groups accordingly Image areas (face areas) are taken into account. Likewise, the Pixel groups in the center of the video have a higher priority Experienced.

Another possibility for optimization is in the fact that neighboring groups of pixels overlap each other. By skilfully selecting the pixel groups it can be avoided that overlapping Pixel values neighboring pixel groups are repeatedly transmitted.

The effort used for decoding is freely scalable. For smaller displays (e.g. cell phones) less computing effort than playback is certainly necessary on a high resolution Large screen, although both have the same source data stream consisting of the prioritized ones Use pixel groups. This flexible scaling allows manufacturers of end devices, special optimizations, e.g. Number of objects, history of image changes, in their devices install. For The manufacturer gives a way to stand out from their competitors stand out without compatibility the video transmission to endanger.

Claims (14)

Method for compressing video data consisting of an array of individual picture elements (pixels), each pixel ( 0-419 ) has a temporally changing pixel value that describes the color or brightness information of the pixel, characterized by the steps: a) determining a priority value for each pixel of the array by calculating a pixel difference value on the basis of the respective current pixel value of the pixel in relation to the current pixel values of one predetermined group of neighboring pixels; b) combining the pixels used for the calculation of the priority value to form a pixel group (P0-P4), c) sorting the pixel groups on the basis of their priority value and storing them in a priority array; and d) storing and / or transferring the pixel groups according to their priority in the priority array, steps a) to d) being repeated continuously, the priority values of the pixel groups always being determined anew and the priority array containing the pixel groups sorted according to current priorities at all times , A method according to claim 1, characterized in that the pixel difference value is the difference of the pixel value of a viewed pixel to the pixel value of each of its viewed Neighbor pixel of the pixel group results. Method according to one of the preceding claims, characterized characterized that at the beginning the parameters of the video picture, such as picture width in pixels, image height stored and / or transmitted in pixels and in the form of the pixel group used become. Method according to one of the preceding claims, characterized characterized that for each pixel group the position of a reference pixel (P0), its pixel value, as well as the pixel value of the others Pixels (P1-P4) of Pixel group saved or transferred becomes. Method according to one of the preceding claims, characterized characterized in that pixel groups of certain image areas are assigned a higher priority becomes. Method according to one of the preceding claims, characterized characterized that the pixel values of the pixel groups by run length coding or other compression methods can be compressed further. Method according to one of the preceding claims, characterized characterized that the ongoing identification and issue of the after priorities sorted pixel groups already by a used image-taking System such as Scanner, CCD camera. Method according to one of the preceding claims, characterized marked that frame grabber cards (or software solutions) can be used to also present video material of various formats (e.g. AVI, MPEG-1, MPEG-2, MPEG-4, ...) to convert. Process for the reconstruction of video data with the method according to claims 1 to 8 were compressed, characterized in that the respectively read Pixel values are displayed in the form of an image array, the not yet transferring Pixels from those already transmitting Pixels are calculated. A method according to claim 9, characterized by the steps: a) Generate an empty image array from the read parameters of the compressed video image, b) Continuous reading of the stored or transferred Pixel groups and insertion into the image array, c) Form triangles by connecting three immediately adjacent pixel groups by at least a line, d) Fill in of the area of the triangles forming one of the pixels forming the triangle Pixel groups calculated color and / or brightness curve, and e) Repeating steps b) to e) Method according to one of claims 9 or 10, characterized in that that the triangles are scalable in size and different image resolutions are customizable. Method according to one of claims 9 to 11, characterized in that that additional Arrays can be created the information contains about the: - Time, i.e. when was a pixel value last calculated or transferred, - basis of calculation, i.e. which transferred Pixels were used to calculate the pixel value, - probability / accuracy, i.e. a pixel value was transmitted or calculated; if it was calculated, how big is it Variance of the pixel groups from which the new value was calculated, - deviation of already calculated pixel values with then transferring pixel values. Method according to one of claims 9 to 12, characterized in that that based on the claim 12 generated arrays, simple movement profiles and objects can be recognized. A method according to claim 13, characterized in that by evaluating the motion profiles and objects smooth movements with extremely low latency, despite the lowest transfer rates can be achieved.